Extensive conflict analysis of data mining based on evidence theory

Abstract

On the basis of combining correlation function with belief function, this paper establishes the analysis and processing mechanism of contradictions form extension association and similarity. In addition, support and believability, conflict and extensibility of data under uncertainty are built through the union of Extenics and data mining. At last, we discuss the conflict transformation conditions, changes space and optimum selection of collision problem based on extension conversion in order to improve the accuracy and the reliability of conflict resolution and rule-based reasoning of complex relationship problems in extension data mining. © 2011 Springer-Verlag Berlin Heidelberg.